Banking Frontiers

Key advantages for data-driven enterprise­s


Aresearch study on the use of data by enterprise­s by 451 Research lists 5 trends that separate enterprise­s that are more data-driven from the least data-driven. The study lists these trends as:

1. The future of AI will be (more)

automated and explainabl­e.

The shortage of data science skills means that unless machine learning gets easier to implement, progress and adoption will be slow. The long-term goal must be to open up machine learning to all types of developers, not just those with strong data science skills. This also means eventually opening up machine learning to users of l ow-code and no-code developmen­t environmen­ts.

2. The ‘ last mile’ of data science - operationa­lization - will be addressed.

Offerings for placing AI and machine learning models into production and handling their ongoing management will burgeon in order to tackle this critical part of data science.

3. DataOps will drive more agile data


To become more datadriven, enterprise­s need to invest not only in new data processing, analytics and machine learning functional­ity, but also in more agile approaches to data management that reduce data friction and accelerate time to insight.

4. Hadoop and dat a warehouse

workload convergenc­e will increase.

When Apache Hadoop first appeared on the scene in 2006, it was more a complement than a competitor to existing data warehousin­g systems. However, distribute­d data-processing frameworks vendors (initially Hadoop distributo­rs) have changed since then and are coming to market with specific data-warehousin­g offerings and moving beyond serving as complement­ary systems to become direct competitor­s to traditiona­l data warehousin­g vendors.

5. Industry-specific and cross-industry

enterprise blockchain­s will go live.

The blockchain crowd is not getting any smaller, but the good news is that more practical uses of the technology are taking off and we are seeing a growing number of industry-specific and crossindus­try enterprise blockchain networks going live.

The study maintains that the pace of change will continue to accelerate in 2019 and that these trends are actually influencin­g each other to drive that change.


The study elaborates on the winners and losers because of the each of the trends. It says in the case of the first trend, the winners will be those that automate the training of machine learning models; data-driven organizati­ons that understand that large disparate data sets are assets to be embraced and exploited. However, organizati­ons that rely heavily on expensive data science skills or insist on building their own models by hand and vendors that only offer machine learning services via APIs, will stand to lose.

In the case of the second trend, winners will be companies that acknowledg­e the role of AI and machine learning operationa­lization in making pervasive intelligen­ce effective and successful and vendors who make machine learning and AI operationa­lization a core feature of their data science platform, cloud service or data platform. And the losers will be those companies that fail to respond to the importance of operationa­lization and vendors with a limited view of a data science pipeline.

As regards DataOps and more agile data management, the study says vendors who deliver products and services that are specifical­ly designed to support more agile data management processes and lower time to insight and enterprise­s that adopt organizati­onal and cultural change to break down barriers between IT and line of business will stand to gain, while vendors that are locked into traditiona­l approaches to data management and enterprise­s that are too conservati­ve to embrace new approaches to data management and analytics will be the losers.

On the aspect of Hadoop and data warehouse workload convergenc­e increasing, the study feels companies that make better use of their Hadoop systems and those that avoid jumping to conclusion­s will be the winners, while those that rush to dismiss the distribute­d data processing systems and organizati­ons that do not fully understand what is required will be the losers.

Pertaining to the fifth aspect of industrysp­ecific and cross-industry enterprise blockchain­s going live, the study emphasizes that vendors with a hands-on approach who can articulate the capabiliti­es and benefits of their offering and enterprise­s that are open to new ways of doing business will be the winners, while the losers will be those vendors who develop their offerings in isolation and fail to implement a sustainabl­e business model and enterprise­s that are too conservati­ve to embrace new and disruptive approaches to doing business.

Newspapers in English

Newspapers from India